• How does R plot maps?
• What is a straight line, anyway?
• Large spatial datasets with R
• Lifecycle of R Spatial packages
• R Spatial community

• Unless you are in the business of actually creating globes,

… any two-dimensional plot of Earth-bound surface data involves projection

plot(map)

maps::map("world")
maps::map("usa")

?map
projection: character string that names a map projection to use.  See
          ‘mapproject’ (in the ‘mapproj’ library).  The default is to
          use a rectangular projection with the aspect ratio chosen so
          that longitude and latitude scales are equivalent at the
          center of the picture.

https://xkcd.com/977/

https://xkcd.com/977/

Projection 101

• For small area’s, projection is usually no big issue (Earth is nearly flat)
• Projections cannot preserve distances; only distances to a single location
• Projections preserve area or shape or direction or some compromise, or …
• There are many of them!

• there is no need for
  • North to be up
  • Europe to be in the middle
• but a projection of a randomly rotated Earth is often harder to read!

For simple features, some definitions:

• feature: abstraction of real world phenomena, with geometrical properties
• simple feature: feature with all geometric attributes described piecewise by straight line or planar interpolation between sets of points

Straight lines, after (re)projection, are no longer straight.

GeoJSON IETF standard:

• A line between two positions is a straight Cartesian line, the shortest line between those two points in the coordinate reference system. […]
• The coordinate reference system for all GeoJSON coordinates is a geographic coordinate reference system, using the World Geodetic System 1984 (WGS 84) [WGS84] datum, with longitude and latitude units of decimal degrees.

• but do GeoJSON users realise that?

This example is contrived, but relevant for every line except sections of meridians or the equator.

For data with ellipsoidal (long/lat) coordinates, sf 1.0-0 switched to using spherical geometry (\(S^2\)) rather than Cartesian geometry (\(R^2\)). Straight lines are now great circle segments on a sphere.

As the Earth’s shape is closer to a sphere than to a flat plane, this is a good thing, but surprises are going to show up for a while due to doing it wrong for some 50+ years.

To get the “old” (pre-sf 1.0) behaviour, use

• sf_use_s2(FALSE),
• st_set_crs(NA), or
• project to to +proj=eqc.

More discussion on this in the (upcoming) Spatial Data Science book by Roger Bivand and me; also do follow Dewey Dunnington

• too large to
  • hold in memory (raster, terra, stars, sf, …)
  • keep in local storage
  • download? (weather: ERA5, climate: CMIP6, earth observation: Landsat/Copernicus)
• Google Earth Engine (and similar) don’t allow you to reproduce analysis, independently
• openEO and openEO Platform are a larger initiatives for allowing reproducible (open source) computing on large, cloud-based data archives, using R and other platforms

• R Spatial is an open and friendly community of people using R for handling and analysing spatial data
• It uses and interfaces a lot of software shared with a much larger community (e.g. OSGEO)
• rgdal and rgeos will retire Jan 1, 2024; sf, terra, stars provide alternatives
• sf now uses spherical geometry:
  • straight lines may need noding,
  • simplifying should be done after projection
• analysing large spatial data sets is, and will remain a challenge